Bug#649673: [powerpc] fails to boot - immediate oops

[Jonathan Nieder <jrnieder@gmail.com>]

retitle 649673 [powerpc] immediate oops on boot (Kernel access of bad area, sig: 11)
severity 649673 important
quit

Clea F. Rees wrote:
> On 23 November 2011 03:14, Jonathan Nieder <jrnieder@gmail.com> wrote:
>> cfr wrote:

>>> Kernel oops. Boot failed. Turned machine off at switch. Restarted
>>> choosing to boot the 2.6 kernel at the second prompt.
[...]
> In the meantime, I took a picture when it went oops today. I was
> watching a bit more carefully this time. I get to the login screen and
> then I get the panic.

Perfect.  Transcription, for reference (no usable instruction dump or
backtrace, alas):

|                Oops: Kernel access of bad area, sig: 11 [#1]
|                PowerMac
|                Modules linked in: rfcomm(+) aes_generic cpufreq_stats cpufre
| ufreq_userspace binfmt_misc fuse xt_hl ip6t_rt nf_conntrack_ipv6 nf_defrag_i
| e xt_state ip6table_filter ip6_tables nf_conntrack_netbios_ns nf_conntrack_b
| ag_ipv4 nf_conntrack_ftp nf_conntrack iptable_filter ip_tables x_tables nls_
| luetooth joydev ams appletouch arc4 input_polldev snd_aoa_codec_tas evdev b4
| 80211 ttm snd_aoa_i2sbus snd_pcm drm_kms_helper snd_page_alloc snd_seq snd_t
| nd_aoa_soundbus cfg80211 rfkill ext4 mbcache jbd2 crc16 sg sd_mod sr_mod crc
| echi_hcd scsi_mod firewire_ohci sungem firewire_core ssb usbcore sungem_phy 
| : scsi_wait_scan]
|                NIP: 00000000 LR: 00000000 CTR: 00000000
|                REGS: f439bee0 TRAP: 0000   Not tainted (3.0.0-1-powerpc)
|                MSR: 00000000 <>  CR: 00000000  XER: 00000000
|                TASK = e4c05130[5113] 'modprobe' THREAD: e4c38000
[...]
|                Call Trace:
|                Instruction dump:
|                XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
|                XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX 60000000 XXXXXXXX XXXXXXXX
|                Oops: nonrecoverable exception, sig: 9 [#2]
|                PowerMac

> The attached is the best of the screenshots I
> took. I'm not sure why it says the kernel is 3.0.0-1? If that's what
> it is saying?

Yes, it says it is booted into a v3.0.y kernel.  Maybe your bootloader
didn't notice the updated installed image.

> I take it the idea with the serial/net console to send the messages
> from the kernel during boot to another device which will record them
> instead of/as well as to the screen. So in the case of the net
> console, that would be another machine on the network?

Exactly.

Thanks again.
Jonathan